Game implementation method and system based on AR technology

Abstract

The application relates to the field of AR augmented reality technology, in particular to a game implementation method and system based on AR technology, which comprises the following steps: acquiring scene images and posture data of a user wearing equipment in the process of game implementation; determining whether to adjust filter parameters in scene image filtering; acquiring trajectory difference values of motion interference influence of each image block in the scene images; combining average levels of motion vector similarity of each image block in the current moment scene images and all other image blocks in the previous moment scene images, acquiring characteristic values of the motion interference influence, clustering the image blocks in the current moment scene images, acquiring adjustment coefficients of filter adjustment of all image blocks in the clustering clusters, and adjusting the filter parameters; after the scene images are filtered and enhanced, the virtual reality scene fusion is completed through coordinate conversion, and the game implementation based on the AR technology is completed. The application improves the quality and effect of the game implementation based on the AR technology.

Description

Technical Field

[0001] This application relates to the field of AR (Augmented Reality) technology, specifically to game implementation methods and systems based on AR technology. Background Technology

[0002] In this era of rapid technological development, advancements in hardware and software technologies have provided strong support for the rise of AR games. The emergence of AR development platforms such as Unity, ARKit (for iOS) or ARCore (for Android) has provided a wealth of APIs and tools, enabling developers to easily implement complex functions such as scene recognition, virtual rendering, and interaction design. This has significantly lowered the development threshold for AR games and promoted their diversified development.

[0003] However, in real-world scene recognition and tracking, complex environmental factors significantly interfere with the process. Changes in lighting can cause images to be too bright, too dark, or blurry, while occlusion can lead to missing scene information, affecting the accuracy and stability of recognition and tracking. Simultaneously, the uncertainty of user actions adds to the difficulty; different ranges of motion, speeds, and habits result in complex changes in device posture and position. Traditional image processing technologies have limitations when handling complex scene images, failing to effectively address issues such as noise and edge blurring, making it difficult to meet the high image quality requirements of AR games. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a game implementation method and system based on AR technology, and the specific technical solution adopted is as follows:

[0005] In a first aspect, embodiments of this application provide a game implementation method based on AR technology, including the following steps:

[0006] During the game implementation process, scene images and user-worn device posture data are acquired.

[0007] For each type of pose data, a prediction is made, the difference between each pose data and its predicted value at each time step is analyzed, and the change threshold of each pose data is determined by combining historical user pose data to determine whether to adjust the filtering parameters when filtering the scene image.

[0008] If the filtering parameters need to be adjusted, extract the matching error and motion vector of each image block in the adjacent scene images. Based on the degree of difference between each image block in the current scene image and each of its nearest neighbor image blocks regarding the motion vector, and combined with the average level of the matching error of each of the nearest neighbor image blocks in the continuous scene images, obtain the trajectory difference value of the motion interference effect of each image block in the current scene image.

[0009] By combining the average level of motion vector similarity between each image patch in the current scene image and all other image patches in the previous scene image, as well as the trajectory difference value corresponding to each image patch, the feature value of motion interference influence of each image patch is obtained. By using the matching error between each image patch in the current time and the corresponding image patch in the previous time and the feature value of each image patch, the image patches in the current scene image are clustered. Based on the feature value of each image patch in the cluster and the distance relationship between each image patch and other image patches, the adjustment coefficient of filtering adjustment for all image patches in the cluster is obtained. Based on the adjustment coefficient of each cluster and the preset initial filtering parameter value, the filtering parameters of the scene image are adjusted.

[0010] After filtering and enhancing the scene image, feature points of the scene image are extracted and matched for recognition. Virtual elements and real-world scenes are then merged through coordinate transformation to complete the game implementation based on AR technology.

[0011] Preferably, the method for determining the change threshold of the various posture data is as follows: calculate the average of the absolute values ​​of the differences between the same posture data of all historical users and the current user in the same game process, and use it as the change threshold of the same posture data.

[0012] Preferably, the filtering parameters for determining whether to adjust scene image filtering further include:

[0013] For each time point, if the absolute value of the difference between any pose data and its predicted value exceeds the change threshold of any pose data, the filtering parameters for scene image filtering at the corresponding time point need to be adjusted; otherwise, the filtering parameters for scene image filtering are the preset initial filtering parameter values ​​and do not need to be adjusted.

[0014] Preferably, the method for calculating the trajectory difference value affected by motion interference of each image block in the scene image at the current moment is as follows:

[0015] ;in, Indicates the current moment in the scene image. The trajectory difference value affected by motion interference in each image patch; Indicates the current moment in the scene image. The image patch and the first The difference in motion trajectory changes between neighboring image patches; This represents the i-th frame among all adjacent frames of the scene image before the current time. The first image patch The mean of the matching errors of the r-th nearest neighbor image blocks, where t is the number of nearest neighbor image blocks of the r-th image block.

[0016] Preferably, the method for obtaining the difference in motion trajectory changes between each image patch and its nearest neighbor image patches is as follows:

[0017] The image blocks adjacent to each image block are taken as the nearest neighbor image blocks of each image block. The mean DTW distance of the motion vector between each image block and its nearest neighbor image blocks in all adjacent frames of the scene image before the current time is calculated as the difference value of the motion trajectory change between each image block and its nearest neighbor image blocks.

[0018] Preferably, the method for obtaining the feature values ​​of the motion interference effect of each image block is as follows:

[0019] Calculate the cosine similarity between each image patch in the current scene image and each other image patch in the previous scene image with respect to motion vectors, and calculate the mean of all the cosine similarities obtained for each image patch in the current scene as the consistency difference value of the trajectory change of each image patch at the current time.

[0020] The product of the consistency difference value of the trajectory changes of each image block in the scene image at the current moment and the trajectory difference value affected by motion disturbance is used as the feature value of the motion disturbance effect of each image block in the scene image at the current moment.

[0021] Preferably, the method for calculating the adjustment coefficients for filtering all image patches in the cluster is as follows:

[0022] In the formula, Indicates the first The adjustment coefficients for filtering all image patches in each cluster; Indicates the first In the cluster, the th Feature values ​​affected by motion disturbance in each image patch; Indicates the first The number of image patches in each cluster; and They represent the first In the cluster, the th The and the first The interference distribution judgment value of each image patch is calculated by taking the average of the normalized Euclidean distance between the center pixel of each image patch in the cluster and the center pixels of all other image patches.

[0023] Preferably, the adjustment method for adjusting the filtering parameters of the scene image is as follows: ;in, Indicates the first The standard deviation parameter of Gaussian filtering is applied to image patches in each cluster; Indicates the first The adjustment coefficients for filtering all image patches in each cluster; This represents the preset initial standard deviation parameter for Gaussian filtering.

[0024] Secondly, embodiments of this application also provide a game implementation system based on AR technology, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of any of the above-described game implementation methods based on AR technology.

[0025] As can be seen from the above, the game implementation method and system based on AR technology provided in this application have at least the following beneficial effects:

[0026] This application allows users to first select a game type on the client side and send the selection information to the server. The server then selects and sets parameters accordingly. Considering the instability of user operations in different game processes, which can interfere with image processing effects and the alignment of virtual elements with real-world scenes, this application determines a threshold for judging posture changes on the client side based on historical data of the selected game. The filtering process is then optimized based on this threshold. By comparing the current user's predicted posture information with the threshold and considering the differences in motion changes between adjacent frames of scene images within a short period before the current moment, the application optimizes the scene images captured in real-time during user operations. Its advantage lies in achieving more accurate scene image optimization and filtering processing during different users' operations on the same game.

[0027] By optimizing filtering, the alignment between virtual elements and real-world scenes can be improved and corrected, providing higher-quality information about virtual and real-world scenes for subsequent interactive functions. This comprehensively improves the quality and effect of AR-based game implementation, bringing users a smoother and more immersive gaming experience. Attached Figure Description

[0028] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 The flowchart of the AR-based game implementation method provided in this application;

[0030] Figure 2 This is a schematic diagram illustrating the process of adjusting filtering parameters when filtering scene images during game implementation, as provided in this application. Detailed Implementation

[0031] To further illustrate the technical means and effects adopted by this application to achieve the intended inventive purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation methods, structures, features, and effects of the AR-based game implementation method and system proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0032] Unless otherwise specified and limited, terms such as “comprising,” “including,” or any other variations thereof are intended to cover a non-exclusive inclusion, such that a circuit structure, article, or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the article or device that includes said element. Furthermore, the term “and / or” as used herein includes any and all combinations of one or more of the associated listed items. All technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0033] The following description, in conjunction with the accompanying drawings, details the specific solutions for the AR-based game implementation method and system provided in this application.

[0034] Please see Figure 1 It illustrates a flowchart of the steps of a game implementation method based on AR technology provided in an embodiment of this application, including the following steps:

[0035] In this embodiment, the game implementation method based on AR technology includes obtaining and sending game identifiers, server querying and parameter sending, real-world scene recognition and tracking, and interactive function implementation. The specific game implementation process is as follows:

[0036] S1, Game Identifier Acquisition and Transmission: The user terminal presents a game selection interface to the user. After the user selects a target game, the terminal acquires the corresponding game identifier. This process can be implemented using existing technology. To obtain game information associated with the account, such as game progress and item ownership, the terminal can simultaneously send the user's account and password to the server. Subsequently, the terminal sends the game identifier to the server. The game identifier allows for precise differentiation between different games, as well as different rooms and levels within the same game.

[0037] S2, Server Query and Parameter Sending: After receiving the game identifier, the server queries the local database for the corresponding preset game rules; after completing the query, the server sends back the parameters required by the corresponding preset game rules to the user terminal, providing necessary information for subsequent game preparation.

[0038] S3, Real-world Scene Recognition and Tracking: The gaming device's camera captures images of real-world scenes, and the user end uses image processing technology to preprocess these images. Specifically, Gaussian filtering is used to suppress image noise, and histogram equalization is employed to improve image contrast. Simultaneously, the gaming device's gyroscope, accelerometer, and position sensors collect real-time data on the device's attitude and position. This data is processed using sensor fusion algorithms (such as Kalman filtering) to achieve precise tracking of the device's attitude and position.

[0039] This embodiment considers that in practical AR game applications, differences in game difficulty and variations in the range of user movements can significantly impact scene image capture. When user movements change rapidly and significantly, the device's rapid movement within a short period can lead to a decrease in the quality of the captured images. Simultaneously, traditional image denoising and enhancement techniques are susceptible to interference from complex scenes during processing, resulting in poor image processing effects. This not only affects the accuracy of subsequent image recognition but also reduces the alignment between virtual elements and the real-world scene.

[0040] Therefore, in this embodiment, the filtering parameters in the scene image filtering process are adjusted based on the user's actual posture data to optimize scene image quality and reduce the negative impact of game difficulty selection and user action intensity on the game experience. The specific process in this embodiment is as follows:

[0041] For each type of pose data, a prediction is made, the difference between each pose data and its predicted value at each time point is analyzed, and the pose data change threshold is determined by combining historical user pose data to determine whether to adjust the filtering parameters when filtering the scene image.

[0042] Assessing the degree of motion interference based on posture information: Accurate real-world scene recognition and tracking are key to achieving precise alignment between virtual elements and real-world scenes. Posture data from the user's worn device is collected via sensors; in this embodiment, the posture data includes acceleration, angle, and position information. A time-series prediction algorithm is used to obtain the predicted value for each type of posture data. The absolute value of the difference between the current posture data and its predicted value is calculated and compared with the change thresholds for various posture data.

[0043] For each moment, if the absolute value of the difference between any pose data and its predicted value exceeds the change threshold of that pose data, it indicates a significant variation in user operation pose, increasing interference with the image. In this case, the filtering parameters for the scene image filtering at that moment need to be adjusted for optimization. Otherwise, no adjustment is needed. When filtering the scene image, Gaussian filtering uses a preset initial standard deviation parameter. In this embodiment, the initial standard deviation parameter is [parameter value missing]. It should be noted that, in this embodiment, preferably, the method for determining the change threshold of each type of posture data is as follows: for the same game, calculate the average of the absolute values ​​of the differences between all historical users and the current user in the same game process for the same type of posture data, and use this average as the change threshold of the same type of posture data.

[0044] By using the above methods, the user's posture changes can be analyzed more accurately during the game, which facilitates the subsequent accurate analysis of image quality problems caused by user operations, and allows for the optimization and adjustment of filtering parameters during image denoising and enhancement.

[0045] First, for scene images where filtering parameters need adjustment, it is necessary to analyze the changing characteristics of the currently acquired scene image data and optimize the filtering parameters in image processing accordingly. Specifically, the matching error and motion vector of each image patch in adjacent scene images are extracted. Preferably, in this embodiment, the current time and adjacent previous... Scene images at each time point are used as input in chronological order of acquisition. A block-matching-based motion estimation algorithm is employed to obtain the matching error and motion vector data of image blocks in adjacent scene images. In this embodiment, to improve processing efficiency, a diamond search algorithm is selected for motion estimation over a short period of time, and the value of n is 10.

[0046] For each image block in the scene image at the current moment, the image blocks adjacent to each image block are taken as the nearest neighbor image blocks of each image block. The average DTW distance of the motion vector between each image block and its nearest neighbor image blocks in all adjacent frames of the scene images before the current moment is calculated as the difference value of the motion trajectory change between each image block and its nearest neighbor image blocks in the scene image at the current moment. It should be noted that the number of scene images before the current moment is n, which is 10.

[0047] To further improve the accuracy of the analysis of motion trajectory changes between neighboring image patches, for each neighboring image patch of each image patch, the th nearest neighbor image patch in all adjacent frames of the scene image before the current time is calculated. The first image patch The normalized result of the mean matching error of the nearest neighbor image blocks is obtained by using the Softmax function in this embodiment. The larger the normalized result is, the smaller the impact of the matching error on the comparative analysis of the motion change features of the local area during the user's actual operation.

[0048] Furthermore, based on the differences in the trajectory changes of each image block compared to other neighboring regions during actual user operation, and the average level of matching errors of each neighboring image block in multiple adjacent frames, the trajectory difference value affected by motion interference is calculated, and the calculation formula is as follows:

[0049] ;in, Indicates the current moment in the scene image. The trajectory difference value affected by motion interference in each image patch; Indicates the current moment in the scene image. The image patch and the first The difference in motion trajectory changes between neighboring image patches; This represents the i-th frame among all adjacent frames of the scene image before the current time. The first image patch The mean of the matching error of the r-th nearest neighbor image patch is given. In this embodiment, the number of scene images in the previous multiple frames is n, which is 10, and t is the number of nearest neighbor image patches of the r-th image patch. The smaller the calculated trajectory difference value, the better the matching error of the r-th nearest neighbor image patch in the scene image acquired at the current time. The more significant the motion interference in the local region corresponding to each image patch, the more pronounced the impact.

[0050] Furthermore, to accurately determine the difference in orientation judgment between the current local region and other local regions due to motion interference, the cosine similarity of each image block in the current scene image with respect to the motion vector is calculated with respect to the motion vector of each image block in the scene image at the current moment and each other image block in the scene image at the previous moment. The mean of all the cosine similarities corresponding to each image block in the scene at the current moment is used as the consistency difference value of the trajectory change of each image block at the current moment. Furthermore, the product of the consistency difference value of the trajectory change of each image block in the scene image at the current moment and the trajectory difference value affected by motion interference is calculated and used as the feature value of the motion interference effect of each image block in the scene image at the current moment. The larger the feature value, the more significant the influence of motion interference on the current local region during real-time acquisition, considering the differences in trajectory change and orientation consistency caused by the user's posture changes during actual operation.

[0051] Based on the above analysis, the local motion interference impact characteristics of the scene image are analyzed by comparing the changes in the current user's operation posture. The image blocks in the current scene image are clustered using the matching error between each image block at the current moment and the corresponding image block at the previous moment, as well as the feature values ​​of each image block. In this embodiment, preferably, the matching error between each image block in the current scene image and the corresponding image block in the previous frame scene image, along with the motion interference impact feature values ​​of each image block in the current scene image, are used to form data pairs. A two-dimensional Cartesian coordinate system is constructed, where the horizontal axis represents the matching error and the vertical axis represents the motion interference impact feature values. All image block data pairs are mapped to this coordinate system, and the density peak clustering algorithm is used to cluster the dataset. The purpose is to make the motion interference feature analysis more accurate for image blocks with similar and small matching errors, so as to make targeted adjustments to local areas with similar interference features. It should be noted that the specific clustering process is existing technology and will not be described in detail in this embodiment. Furthermore, other existing clustering methods can be used by the implementer of the clustering method, and this embodiment does not impose any special restrictions on this.

[0052] For each cluster after clustering, the filter adjustment coefficient with similar motion interference characteristics in the cluster is determined based on the distribution characteristics of different local regions in the image patch. The specific calculation method is as follows: calculate the normalized Euclidean distance between the center pixels of each image patch in the cluster and the center pixels of other image patches. The average of all normalized distances obtained for each image patch is used as the interference distribution judgment value for each image patch in the cluster. The more clustered the distribution, the greater the impact of motion interference on similar areas in the user's actual operation, and the greater the impact on filter adjustment. The calculation formula for the filter adjustment coefficient in each cluster is:

[0053] ;

[0054] in, Indicates the first The adjustment coefficients for filtering all image patches in each cluster; Indicates the first In the cluster, the th Feature values ​​affected by motion disturbance in each image patch; and They represent the first In the cluster, the th The and the first The judgment value of interference distribution of each image patch; Indicates the first The number of image patches in each cluster. The larger the calculated filter adjustment coefficient, the greater the impact of motion disturbance on the local area where all image patches in the current cluster are located.

[0055] Finally, based on the filtering adjustment coefficients for different clusters, the filtering parameters for all image patches in each cluster are optimized, and the adjustment relationship is as follows: ;in Indicates the first The standard deviation parameter of Gaussian filtering is applied to image patches in each cluster; The initial standard deviation parameter represents the preset standard deviation parameter for Gaussian filtering. Considering this as a preprocessing step in the game implementation, in this embodiment, the initial standard deviation parameter is set to [value missing]. The size of the Gaussian filter convolution kernel is The implementer can set it up themselves.

[0056] Thus, according to the above process in this embodiment, Gaussian filtering can be performed on each image block by using the standard deviation parameter of the Gaussian filtering in each cluster. At the same time, in this embodiment, the histogram equalization algorithm is preferably used to enhance the filtered scene image.

[0057] After the above filtering and enhancement processes, feature points of the scene image are extracted and matched for recognition. Virtual reality scene fusion through coordinate transformation completes the AR-based game implementation. The specific process can be implemented using existing technologies, and this embodiment does not impose any special limitations on it. In this embodiment, the ORB (Oriented Fast and Rotated BRIEF) algorithm is used to extract feature points in the scene image, and a Hamming distance-based matching algorithm is used to identify specific objects, planes, or feature points in the scene. Based on the accurate tracking results of the device's posture and position, a coordinate transformation algorithm is used to transform the coordinates of the virtual elements to the real-world scene coordinate system, ensuring accurate alignment between the virtual elements and the real-world scene, and adjusting their position and angle in real time as the device moves. The user terminal, based on the real-world scene recognition results and device tracking information, appropriately places the virtual elements and merges them with the real-world scene.

[0058] Specifically, in this embodiment, the flowchart illustrating the adjustment of filtering parameters during scene image filtering in the game implementation process is as follows: Figure 2 As shown.

[0059] S4, Interactive Function Implementation: Implement touch interaction functionality, allowing users to click virtual elements and control character movement via the touch screen. During gameplay, the user terminal sends the game level identifier and game result to the server. The server determines whether a match is found based on preset passing criteria, which are implemented using existing technology; this embodiment does not impose special restrictions on this. If the game result is a match, the server sends a level-opening command, and the user terminal unlocks the next level; if the game result is a mismatch, a level-closing command is sent. Furthermore, before the game starts, the implementer can also preset criteria to verify external devices. Before the game starts, the server queries the game status based on the user identifier to determine if it meets the preset criteria, thus deciding whether to allow the external device to start the game.

[0060] Based on the same inventive concept as the above methods, this application also provides a game implementation system based on AR technology, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described game implementation methods based on AR technology.

[0061] It is understood that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0062] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0063] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Any equivalent structural or procedural transformations made based on the description and drawings of this application, or direct or indirect applications in other related technical fields, are similarly included within the protection scope of this application.

Claims

1. A game implementation method based on AR technology, characterized in that, Includes the following steps: During the game implementation process, scene images and user-worn device posture data are acquired. For each type of pose data, a prediction is made, the difference between each pose data and its predicted value at each time step is analyzed, and the change threshold of each pose data is determined by combining historical user pose data to determine whether to adjust the filtering parameters when filtering the scene image. If the filtering parameters need to be adjusted, extract the matching error and motion vector of each image block in the adjacent scene images. Based on the degree of difference between each image block in the current scene image and each of its nearest neighbor image blocks regarding the motion vector, and combined with the average level of the matching error of each of the nearest neighbor image blocks in the continuous scene images, obtain the trajectory difference value of the motion interference effect of each image block in the current scene image. By combining the average level of motion vector similarity between each image patch in the current scene image and all other image patches in the previous scene image, as well as the trajectory difference value corresponding to each image patch, the feature value of motion interference influence of each image patch is obtained. By using the matching error between each image patch in the current time and the corresponding image patch in the previous time and the feature value of each image patch, the image patches in the current scene image are clustered. Based on the feature value of each image patch in the cluster and the distance relationship between each image patch and other image patches, the adjustment coefficient of filtering adjustment for all image patches in the cluster is obtained. Based on the adjustment coefficient of each cluster and the preset initial filtering parameter value, the filtering parameters of the scene image are adjusted. After filtering and enhancing the scene image, feature points of the scene image are extracted and matched for recognition. Virtual elements and real-world scenes are then merged through coordinate transformation to complete the game implementation based on AR technology.

2. The game implementation method based on AR technology as described in claim 1, characterized in that, The method for determining the change threshold of various posture data is as follows: calculate the average of the absolute values ​​of the differences between the same posture data of all historical users and the current user in the same game process, and use it as the change threshold of the same posture data.

3. The game implementation method based on AR technology as described in claim 1, characterized in that, The filtering parameters used to determine whether to adjust scene image filtering further include: For each time point, if the absolute value of the difference between any pose data and its predicted value exceeds the change threshold of any pose data, the filtering parameters for scene image filtering at the corresponding time point need to be adjusted; otherwise, the filtering parameters for scene image filtering are the preset initial filtering parameter values ​​and do not need to be adjusted.

4. The game implementation method based on AR technology as described in claim 1, characterized in that, The method for calculating the trajectory difference value affected by motion interference of each image block in the scene image at the current moment is as follows: ;in, Indicates the current moment in the scene image. The trajectory difference value affected by motion interference in each image patch; Indicates the current moment in the scene image. The image patch and the first The difference in motion trajectory changes between neighboring image patches; This represents the i-th frame among all adjacent frames of the scene image before the current time. The first image patch The mean of the matching errors of the r-th nearest neighbor image blocks, where t is the number of nearest neighbor image blocks of the r-th image block.

5. The game implementation method based on AR technology as described in claim 4, characterized in that, The method for obtaining the difference in motion trajectory changes between each image patch and its nearest neighbor image patches is as follows: The image blocks adjacent to each image block are taken as the nearest neighbor image blocks of each image block. The mean DTW distance of the motion vector between each image block and its nearest neighbor image blocks in all adjacent frames of the scene image before the current time is calculated as the difference value of the motion trajectory change between each image block and its nearest neighbor image blocks.

6. The game implementation method based on AR technology as described in claim 1, characterized in that, The method for obtaining the feature values ​​of the motion interference effect of each image block is as follows: Calculate the cosine similarity between each image patch in the current scene image and each other image patch in the previous scene image with respect to motion vectors, and calculate the mean of all the cosine similarities obtained for each image patch in the current scene as the consistency difference value of the trajectory change of each image patch at the current time. The product of the consistency difference value of the trajectory changes of each image block in the scene image at the current moment and the trajectory difference value affected by motion disturbance is used as the feature value of the motion disturbance effect of each image block in the scene image at the current moment.

7. The game implementation method based on AR technology as described in claim 1, characterized in that, The method for calculating the adjustment coefficients of the filtering adjustment of all image patches in the cluster is as follows: In the formula, Indicates the first The adjustment coefficients for filtering all image patches in each cluster; Indicates the first In the cluster, the th Feature values ​​affected by motion disturbance in each image patch; Indicates the first The number of image patches in each cluster; and They represent the first In the cluster, the th The and the first The interference distribution judgment value of each image patch is calculated by taking the average of the normalized Euclidean distance between the center pixel of each image patch in the cluster and the center pixels of all other image patches.

8. The game implementation method based on AR technology as described in claim 1, characterized in that, The method for adjusting the filtering parameters of the scene image is as follows: ;in, Indicates the first The standard deviation parameter of Gaussian filtering is applied to image patches in each cluster; Indicates the first The adjustment coefficients for filtering all image patches in each cluster; This represents the preset initial standard deviation parameter for Gaussian filtering.

9. A game implementation system based on AR technology, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the AR-based game implementation method as described in any one of claims 1-8.

Images